Tuesday, November 8, 2016

Digital voice stuff

It's been a while since I blogged on that area.  There have been a number of proposed multi-protocol digital radios, and I guess I have been holding out for that.

And while KN4AQ's reddit post is a pretty good summary of the issues on why that is taking so long to come to be, it doesn't really touch upon the digital voice software experimentation and progress being worked up independently by hams.

Before I dive into that more, some commentary on Yaesu System Fusion.  This was the later player to step to the plate in 2013.  And while Yaesu has pretty much been my radio of choice over the years, I guess I had higher hopes figuring they learned a few things along the way, getting into it later.

Some things they did are done right in my opinion.  At least their repeaters are backwards compatible with analog.  It's a good marketing strategy for them.  But mostly one doesn't have to launch another repeater specific to one mode.  So it's a prudent move in terms of spectral efficiency.  I mean don't we have enough idle repeaters already?

The whole external Wires X, controller and computer to provide external internet connectivity is horrid to me.  I guess I was hoping for a repeater with a network jack, like DMR.  I'd take that over all the touch screen sillyness.

On the user end of Yaesu System Fusion radios I am glad they are firmware updatable? (I know the repeater is for sure).  It's really quite sad and dumb that the firmware is a closed deal, not open to 3rd party development. And the whole camera thing is really dumb in my opinion.

What I have decided to monkey with is MMDVM, which is a open-source Multi-Mode Digital Voice Modem project by G4KLX.   It allows one to retro fit an existing analog repeater into one that supports, D-Star, Yaesu System Fusion, DMR and P25.  I have my name down for the interface board and expect to be playing with this project soon.

I was thinking about the DV4mini and some of the other stuff like the DVmega, but I am turned off by the low power/ single user limitations.  I have always been a fan of building/buying things a number of folks can use... the community approach.

Prior to DMR becoming popular in ham radio, and Yaesu introducing their Fusion stuff, Codec 2 made some sense.  At that point it was pretty much just D-Star, and everyone was in a tizzy about it using a propriety codec/vocoder.  Now there are so many different VHF/UHF digital radios out there using the patented technology, and there still isn't a drop in "open" replacement for VHF/UHF.  Since all the major amateur radio players have their own digital radios, even if an open codec were available tomorrow, it would take a start-up company to adopt it and and a miracle to come up with a priced right radio that would potentially displace the existing digital market in the foreseeable future because as as noted there been a number of setbacks in developing multi-protocol user radios.  And in one year the open source AMBE code for D-Star should be legal.

So that is where we sit for now, and probably quite a while.  Sad, but true.

A couple years ago Bruce Perens gave a pretty good history/background speech titled "AMBE Exposed"  What we need now are some specifics (a follow up talk).  He talked about trying to invalidate the patent based on David Rowe's work and that DVSI made use of AMBE in commerce before their own patent applications.

Presently Jonathan's MMDVM (mutli-mode digital voice modem) repeater project has the capability to tie things together using multi-mode conference bridges.  We have heard a little bit about a the CCS7 reflector system that is popular in Europe, that does just that to some extent too.

DMR, System Fusion DN, and NXDN all use exactly the same bit rate and FEC. "It's simply a matter of unwrapping the AMBE stream from whichever container format is used and wrapping it in the target format."  Anything else may require hardware or a much better understanding of the patents.

Presently all the hardware dongles to deal with AMBE can only process a single stream at a time.  This is less than ideal for anyone interesting is hosting a multimode conference reflector system.

The patent for IMBE for P25 may have expired.  That is one thing that would be nice to get specifics on, as none of the hardware dongles really support that mode.

If someone were to start unwrapping the older AMBE that D-Star uses to create a software bridge to the newer AMBE+2 used in DMR, would that perhaps be permitted?

What algorithm is exactly protected by the patent?  Obviously taking it all the way back to analog is a patented thing.


Which patents apply to IMBE, which to AMBE+, and which to AMBE+2?

When they expire how will be know for sure?  If a commercial vendor were to seek the licensed code, is the patent holder under any obligation to inform the vendor its out of patent, and they may freely implement their own?

An example case of some other formerly patented technology, like Motorola's Private Line, or the like, and how generics came about might be a good explanation to many.

Note:  For most U.N. member states; non-commercial/research usage of patented technology is covered by exceptions on the definition of "patent infringement."  Actual practice may differ. Ref